Skip to content

CLI Reference

Jump to bottom
Calvin Yang edited this page Jun 1, 2026 · 1 revision

CLI Reference

python grr_tool.py [options]

Core flags

Flag Short Default Description
--input -i required Input CSV (Part, Operator, Trial1, Trial2, ...)
--output -o auto PDF report output path
--dashboard -d none HTML dashboard output path
--title GR&R Study Report Report/dashboard title
--equipment -e Unspecified Gage Equipment identifier
--operator Quality Engineering Name of QE or team
--generate-sample off Generate synthetic 10x3x3 CSV before analysis
--version -v Show version and exit

Tolerance flags

Flag Description
--tolerance Full engineering tolerance range (e.g. 0.050 for +/-0.025 spec)
--usl Upper spec limit — use with --lsl for asymmetric specs
--lsl Lower spec limit — use with --usl

Tolerance from --usl/--lsl = USL - LSL. Also computes ndc_tol = 1.41 * (tolerance / GRR).

Study type

--study-type {crossed,nested,linearity,bias}
Value Description
crossed Standard AIAG crossed GR&R (default)
nested Nested/destructive GR&R — each operator measures unique parts
linearity Linearity study — requires --reference-values
bias Bias study — requires --reference-values (single value)

--reference-values

Comma-separated known reference values (one per unique part for linearity; one value for bias).

# Linearity (10 parts)
python grr_tool.py --input data.csv --study-type linearity \
  --reference-values "9.97,9.975,9.98,9.985,9.99,9.995,10.0,10.005,10.01,10.015" \
  --tolerance 0.060

# Bias
python grr_tool.py --input data.csv --study-type bias \
  --reference-values "10.000" --tolerance 0.060

Attribute flag

--attribute

Switches to Attribute Agreement Analysis. Trial values must be 0 or 1.

Computes: within-operator agreement %, between-operator agreement %, kappa per operator, and effectiveness % (requires --reference-csv).

Effectiveness Verdict
>= 90% Acceptable
80-90% Marginal
< 80% Unacceptable

--reference-csv

Two-column CSV (Part, Reference) with known correct result (0 or 1) per part.

Run chart

--run-order COLUMN_NAME

Enables time-series run chart and Nelson rules 1, 2, 5, 6 detection.

Rule Description
1 Any point beyond 3 sigma
2 9 consecutive points on same side of mean
5 2 of 3 consecutive points beyond 2 sigma (same side)
6 4 of 5 consecutive points beyond 1 sigma (same side)

Multi-study comparison

--compare file1.csv file2.csv ...

Generates a PDF with side-by-side table and bar charts for %GRR, ndc, EV%, AV%, PV%.

python grr_tool.py \
  --compare before.csv after.csv --output improvement.pdf

Examples

# Standard study — PDF + dashboard with tolerance
python grr_tool.py --input data.csv --output report.pdf --dashboard dash.html \
  --equipment "Mitutoyo 293-340-30 SN-1234" --tolerance 0.050

# Asymmetric tolerance
python grr_tool.py --input data.csv --output report.pdf --usl 10.025 --lsl 9.975

# Nested / destructive
python grr_tool.py --input destructive.csv --study-type nested

# Attribute go/no-go
python grr_tool.py --input gonogo.csv --attribute --reference-csv known.csv

# Quarterly comparison
python grr_tool.py --compare q1.csv q2.csv q3.csv --output quarterly.pdf

Clone this wiki locally